[bitc-dev] Issues for s-block surface syntax

Sandro Magi naasking at higherlogics.com
Fri Jan 25 17:49:36 EST 2008


Jonathan S. Shapiro wrote:
> Sandro: I agree, but note you are now talking about introducing a very
> dangerous language feature in order to support .000001% of programmers.  

Operators are useful mainly to the end-users, not to the library developers, so the market is rather larger than that. Unless you're referring to functional languages being used by 0.00...1% of programmers?

>> So a facility is broken because it doesn't help 0.001% of the market reviewing code on paper? I don't find that convincing.  
>
> I do. The fundamental and primary driving objective of BitC is to make
> code robust and comprehensible (both in the human sense and the machine
> sense). If that doesn't match your priorities, that is okay, but the
> language you want isn't BitC.  

I wasn't speaking to BitC specifically, just for general purpose languages with flexible syntax in general. I would prefer flexible operators over using PEGs or packrat parsing with *fully* customizable syntax (as some languages are adopting). I understand that BitC might target a different market entirely.

> That being said, I don't reject the idea of better tools. But ML has had
> this feature for decades, and I don't see any tools yet.  

There also aren't any IDEs, which I would think is a prerequisite for such a tool. It's a shame too. F# is the closest ML-derivative to having an IDE.

Sandro






More information about the bitc-dev mailing list